© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 24280
    Prev
    Next
    INDEX
    Explanations

    phrases containing conditional or descriptive clauses

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    109
    +0.11
    0.5%
    757
    +0.11
    0.5%
    420
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    655
    +0.11
    0.07
    109
    +0.11
    0.06
    757
    +0.10
    0.06
    Negative Logits
     Ideas
    -0.67
     Planning
    -0.67
    ONE
    -0.61
     Equity
    -0.61
     Doctrine
    -0.60
    ODY
    -0.60
    Politics
    -0.59
     minded
    -0.59
     Thoughts
    -0.58
     arts
    -0.58
    POSITIVE LOGITS
     resembles
    1.19
     corresponds
    1.06
     resemble
    1.04
     protr
    1.04
     weighs
    1.02
     translates
    0.99
     resembled
    0.98
     contains
    0.94
     circulate
    0.91
     protects
    0.90
    Activations Density 0.203%

    No Known Activations